Does the Basel III countercyclical capital buffer mitigate regulatory capital (pro)cyclicality? Evidence from empirical tests - preliminary and incomplete Massimo Libertucci Banca d’Italia joint work with Ulrich Krüger (Deutsche Bundesbank) EBA Policy Research Workshop London, 14 November 2013
Disclaimer The opinions expressed are those of the authors and do not necessarily reflect those of Bank of Italy and Deutsche Bundesbank
Outline 1. Introduction – Motivation and main results 2. Stylized facts for financial cycles in Germany and Italy 3. Empirical estimation of firms ´ probabilities of default – Empirical estimations of PDs – Data 4. CCB and procyclicality: an empirical test – Methodology – Results 5. Extensions 6. Conclusions and policy implications
Introduction 1/3 The introduction of the Countercyclical Capital Buffer (hereafter: CCB) constitutes one of the most remarkable innovation within the new Basel III framework • the CCB is thought to play a pre-emptive role, imposing banks to progressively build-up capital cushions as imbalances in credit market develop • the CCB first goal is to shelter banking sector from the consequences of excessive credit growth by increasing banking sector loss-absorbing capacity • additionally, CCB may lean against the build-up of excesses by affecting the cost of credit. The signs of the presence of increasing imbalances will spur the activation of CCB Procyclicality constituted (one of) the reasons for the introduction of CCB • risk-sensitive capital requirements have been associated with a potential drawback in the form of amplifying the cyclical fluctuations of the economy via the lending channel 4
Introduction 2/3 The 2009 consultative document issued by Basel Committee for Banking Supervision (BCBS 2009) delineated – inter alia - a complete toolkit to tackle procyclicality … BCBS 2009 proposal initially incorporated 4 different instruments, each of one aimed at a different policy target. √ conserve capital to build buffers that can be used in stress √ achieve the broader macro-prudential goal of protecting the banking sector from periods of excess credit growth x dampen any excess cyclicality of the minimum capital requirement √ the fourth issue of provisions has been addressed by the Accounting world; thus to some extent it lies behind the cycle of banking supervision The first 2 objects have experienced a follow up within the final draft of the Basel III framework in the form of 2 policy instruments: • the capital conservation buffer and the countercyclical capital buffer respectively 5
Introduction 3/3 Conversely, the proposal initially included in the BCBS (2009) consultative paper to directly tackle excessive cyclicality of capital requirements has known a different fate • BCBC (2009) referred to some specific proposals already put forward in the reform debate (e.g. CEBS 2009 Pillar 2 or UK FSA 2009 non-cyclical PDs) • core of these proposals: an additional capital buffer deriving from the introduction of a bank-specific mechanism that accounts for the historical performance of the PDs during the cycle 6
Motivation and main results 1/3 In spite of all potential gains, a micro-based anti-cyclical tool has been left out the final draft of the proposal • appetite of the regulatory community for avoiding a proliferation of new risk measures potentially prone to the same form of fallacy of composition already existing Commentators have started to shed doubts on the real ability of the anticyclical toolkit in its final design to achieve all its targets • Enria and Quagliariello (2010) stress the importance of introducing a micro-perspective (in their definition, represented by PD-smoothing alike proposals) into the macro-prudential toolkit • Repullo and Saurina (2011) argue that the CCB may end up by exacerbating the inherent pro-cyclical nature of the Basel framework, rather than reducing it 7
Motivation and main results 2/3 The need to further evaluate the actual anti-cyclical potential of the CCB motivates this paper • we follow Repullo, Saurina, and Trucharte, 2010 • we test whether CCB can achieve a secondary policy goal • we assess CCB performance with respect to some of the leading alternative procedures that have been proposed to mitigate the excess cyclicality issue • we do so minimizing the deviations of each adjusted series with respect to a benchmark, in the form of a filtered version of the original capital requirement series 8
Motivation and main results 3/3 Preliminary results show that, under the hypotheses of this study, the CCB performs poorly in tackling excessive volatility of capital requirement • more limited potential whether compared to other options, as a correction factor based on business cycle variables (e.g. GDP growth) We refrain assessing the ability of CCB to achieve its primary broader goal of protecting the banking sector from periods of excess credit growth • a suitable design of CCB can make this job (Drehmann and Juselius, 2013) • some insights of this work may contribute to inform the current debate on implementation of CCB (e.g. ESRB role under CRD IV) 9
Stylized facts for financial cycles in DE and IT 1/3 To better comprehend the functioning of the CCB, the anatomy of the financial cycle represents the natural starting point (Drehmann et al. (2010), Arnold et al. (2012)): three sets of stylized facts • the behavior of private sector credit and property prices provides the best description of financial cycle. On the contrary, equity bubble bursts represent a mere hiccup in the longer financial cycle • the peak of financial cycle tends to coincide with episodes of severe financial distress; conversely, in few episodes these peaks do not end up with a crisis • the basis of the previous two sets is it possible to build up a series of early warning indicators of financial crisis. These indictors hinge on private credit and property prices 10
Stylized facts for financial cycles in DE and IT 2/3 The final design of the CCB (BCBS 2010) fully incorporates the above three facts • The deviation credit-to-GDP gap represents the preferred reference variable for the build-up phase (BCBS). The use of a deviation from a medium-term trend is consistent with the theoretical facts depicted above. A graphical inspection of this variable time series represents a starting point for the analysis 11
Stylized facts for financial cycles in DE and IT 3/3 Some insights drawn from graphical inspection • Type I and Type II errors: starting from the two charts, highlight how the Gap was able to identify some crisis phenomena (e.g. Italy 2008) but missing other (Germany 2008) • Signal of overheating issued (high Gap) even after the explosion of the 2008 crisis in Italy The relation between the credit-to-GDP ratio gap and the business cycle represents a major concern • financial cycle is neither necessarily nor normally in synch with the business cycle • correlation analysis tend to confirm it: negative correlation coefficients this may cause risk of building up buffer in recessionary periods (Repullo et al. 2011) 12
Empirical estimation of PDs 1/2 The core of our analysis aims at evaluating the effects of the interaction of the CCB with some Basel II-alike capital requirements • an adequately long time series of risk sensitive bank capital requirements represents the necessary input of the analysis We estimate models of default probability to use as inputs to calculate capital requirements • we use available information to compute how the capital requirements would have been evolved over time if Basel II had been in order • admittedly, this approach represents a second-best solution, prone to Lucas ´ critique • nevertheless, this procedure shows some desirable features (Kashyap and Stein (2004), Saurina and Trucharte (2007); and CEBS 2009) 13
Empirical estimation of PDs 2/2 We use two different rating models for German and Italian firms • for German firms, we use the model developed by Krüger, Stötzel and Trück (2005) • for Italian firms, we use the rating model developed by Chionsini, Fabi and Laviola (2007) Data sources: Germany • information from a database of annual balance sheet information of non-financial enterprises to classify trade-bills of companies as eligible collateral for the use in refinancing operations. • data quality issues and structural break (2003) led to the decision to split into in two sub-samples: 1988-2001 and 2005-2011 Data source: Italy • the Italian Central Credit Register: credit information record for all the credit transactions with a value above EUR 35,000. The Register contains all the relevant information related with the characteristic of a given loan Table 1 shows most relevant information 14
CCB and procyclicality: an empirical test 1/12 Methodology • starting from the above PDs, we work out the corresponding Basel II Point- In-Time (PIT) IRB capital requirement for each borrower in the sample for a given year, assuming Basel II had been in place • we use the F-IRB corporate regulatory formula • we then calculate the annual averages Average PIT capital requirements are negatively correlated with GDP growth in both countries • the cyclical behavior of risk-sensitive metric implies a significant degree of dispersion of the capital requirements 15
Recommend
More recommend
